Correlates and Prediction of Generalized Anxiety Disorder from Acoustic and Linguistic Features of Impromptu Speech
Bibliographic record
Abstract
The diagnosis and tracking of Generalized Anxiety Disorder (GAD) require frequent interaction with mental health professionals, but there are not enough professionals to serve the demand. The ability to automatically detect anxiety disorders through samples of speech could be a valuable complement to conventional treatment and provide greater access to it. In this work, we explore acoustic and linguistic features of speech that correlate with anxiety and use those to predict above or below the screening threshold of GAD. A large number of participants (N = 2, 000) participated in a single online study session where they completed the Generalized Anxiety Disorder-7 (GAD-7) assessment and provided an impromptu speech sample in response to a modified version of the Trier Social Stress Test (TSST). Acoustic and linguistic speech features were a-priori selected based on the existing speech and anxiety literature, together with related features. Associations between speech features and anxiety levels were assessed using sex, age, and personal income included as covariates. The amount of speech had the most significant correlation with GAD-7 (r = −0.12; P < .001), indicating that participants with higher anxiety scores spoke less. Linguistic features acquired using Linguistic Inquiry and Word Count (LIWC) were also significantly (P < .05) associated with anxiety. Using these acoustic and LIWC features to predict above or below a screening threshold of 10 for GAD, a logistic regression model achieved a mean AUROC = 0.57, SD = 0.03. The mean AUROC increased to 0.62 (SD = 0.03) when demographic information (age, sex, and income) was included indicating the importance of demographics when screening for anxiety disorders. The LIWC linguistic features are based on single-word counts and do not account for the surrounding word context. We attempted the same prediction based on greater context using a transformer-based neural-network model (pre-trained on large textual corpora) and fine-tuned on the speech transcripts. This model, which only uses the textual input, achieved an AUROC value of 0.64. When acoustic, LIWC, and the predicted output of the fine-tuned transformer-based model are combined into one model, the AUROC increased to 0.67 and further increased to 0.68 when demographic information was included. The results suggest that there is a signal of anxiety within such impromptu speech that may be useful as part of a system to screen for anxiety, detect relapse, or monitor treatment.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.005 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.002 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".